Graduated with PhD in ML from Georgia Tech
Thesis titled "On using Inductive Biases for Designing Deep Learning Architectures".
Next Step: Towards becoming an Auror !
Google Research / Google Brain
Research Intern in the Google translate team. My work here is to explore the effectiveness of Randomized Neural Networks for sequence modeling tasks.
GRNUlar: Gene Regulatory Network reconstruction using Unrolled algorithm from Single Cell RNA-Sequencing data
On arxiv (pdf)
Our work on learning Directed Gene Regulatory networks using a novel technique of combining the expressive power of neural networks with unrolled algorithms.
GLAD: Learning Sparse Graph Recovery
Accepted at ICLR-20 conference (pdf)
Joint work with Georgia Tech researchers. A novel unrolled DL algorithm for doing sparse graph recovery. Our work highlights the use of algorithms as inductive bias for DL architecture design.
AntMan: Sparse Low-Rank Compression to Accelerate RNN inference
On arxiv (pdf)
Joint work with Microsoft Research. We were able to compress LSTM based models upto 100X within 1pt loss of accuracy. It is a simple technique and can be easily adopted to existing RNN based models.
2019 (Summer Intern)
X- The Moonshot Factory
I was very fortunate to spend this summer with Alphabet's 'secret' research labs (formerly GoogleX). Researchers there work on futuristic & audacious projects and it was a very memorable experience to contribute to one such endeavour.
2018 (Yet another app...)
2018 (Summer intern)
Microsoft AI Research, Redmond
PhD Research intern (worked on RNN compression, under evaluation in ICLR conference)
Had a great internship experience. Explored the beautiful city of Seattle, visited Alaska, finally did skydiving, met some awesome people. Overall, quite a memorable experience
2017 (Summer intern)
PhD Research intern at Interaction and Analytics Laboratory (IAL)
Worked on anomaly detection for unstructured data for document security. Patented the algorithm developed.
Another Class Imbalance ML classifier published
Accepted at AAAI-2017
Developed a ML classifier based on chi-sq technique. Though, this technique is difficult to scale to higher dimension.
2016 - present
PhD in CS (Machine Learning dept.)
Trying out grad school. Taking time to ponder upon some research ideas.
(Another startup attempt..)
Wiztrit brings Excerpts from the best novels, which make a good 10-minute read, to a single platform. Excerpts have been meticulously screened. These are Comprehensive and Interesting picks from greatest novels of all times. Enjoy browsing through them. Take a glimpse into books you always wanted to read, where just reviews didn’t suffice. Welcome aboard the journey of discovering new books. Happy Reading!!
(Attempt at a startup..)
Android App - Link
The Closelook app links you to all activities and happenings nearby. With its location based service, create or join a circle that interests you and find people sharing similar interests near your location. Follow and interact in the circles to stay updated on your favorite topics. Talk about the maintenance situation in your neighborhood, find best food in your locality, or discuss places to visit in a city.
2014 - 2016
Xerox Research Center India
Budding Scientist, Data Analytics for Healthcare
After graduation from IIT-KGP, I joined Xerox research labs. This was my first job in industrial research setting. As a profit based company, I think they maintained a decent conducive exploratory research environment.
Dilettante --> Professional
Shifted research interest from VLSI to Machine Learning. Started participating in data science competitions. Turns out that playing with probability and optimizations algorithms interests me the most.
TU Munich, Germany
Kudos to German work ethics. It was definitely a Euro trip! haha!
Carnegie Mellon University
Academic commitments aside, had a wonderful stay in Pittsburgh. Certainly was a good opportunity to explore USA.
Indian Institute of Science (IISc), Bangalore
Indian Academy of Science Fellowship
My first academic research experience. IISc has a beautiful campus and clubbed together with incredibly plesant weather, it really has a awesome research feel to it :)
B.Tech (Hons) + M.Tech (Dual degree) - Electronics & ECE
Had a memorable undergrad here!
Copyright @ Harsh Shrivastava. All Rights Reserved